The Agent Skills Directory

[SAFE]: The skill primarily consists of stylistic and structural guidelines for an AI agent to follow when generating images for web design references. It emphasizes quality control, anti-repetitive patterns, and visual hierarchy.
[COMMAND_EXECUTION]: The skill defines a reporting workflow that requires the agent to create and modify files within a local llmwiki directory. This is used for project logging and status updates, representing a legitimate use of file-writing capabilities for documentation purposes.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied design briefs to determine visual parameters. While this involves untrusted input, the skill provides a structured mapping system to interpret the input, and the associated risks are minimal given the skill's primary function of image generation and internal project logging.

imagegen-frontend-web